Intracellular translation of circular RNA

Abstract

A circular mRNA molecule possessing features resembling native mammalian mRNA demonstrates improved translation, while retaining the properties of an extremely long half-life inside cells. This circular mRNA is functional inside mammalian cells, being able to compete against native cellular mRNAs for the eukaryotic translation initiation machinery. The invention possesses additional RNA elements compared to a previous invention containing only an IRES element for successful in vitro or in vivo translation.

Claims

1. A vector for making circular mRNA, said vector comprising the following elements operably connected to each other and arranged in the following sequence: a) an RNA polymerase promoter, b) a self circularizing intron 5′ slice junction, c) an IRES, d) an optional 5′ UTR, e) a multiple cloning insertion site for inserting an ORF into said vector, f) a 3′ UTR, g) an optional polyA tract, h) a self circularizing intron 3′ slice junction, and i) an optional RNA polymerase terminator, said vector allowing production of a circular mRNA that is able to compete against native cellular mRNAs for eukaryotic translation initiation machinery and that is translatable inside eukaryotic cells.

2. The vector of claim 1, wherein the RNA polymerase promoter and terminator are from a T7 virus, T6 virus, SP6 virus, T3 virus, or T4 virus.

3. The vector of claim 1, wherein the 3′ UTR is from human beta globin, human alpha globin xenopus beta globin, xenopus alpha globin, human prolactin, human GAP-43, human eEF1a1, human Tau, human TNF alpha, dengue virus, hantavirus small mRNA, bunyanavirus small mRNA, turnip yellow mosaic virus, hepatitis C virus, rubella virus, tobacco mosaic virus, human IL-8, human actin, human GAPDH, human tubulin, hibiscus chlorotic rinsgpot virus, woodchuck hepatitis virus post translationally regulated element, sindbis virus, turnip crinkle virus, tobacco etch virus or Venezuelan equine encephalitis virus.

4. The vector of claim 1, wherein the 5′ UTR is from human beta globin, Xenopus laevis beta globin, human alpha globin, Xenopus laevis alpha globin, rubella virus, tobacco mosaic virus, mouse Gtx, dengue virus, heat shock protein 70 kDa protein 1A, tobacco alcohol dehydrogenase, tobacco etch virus, turnip crinkle virus, or adenovirus tripartite leader.

5. The vector of claim 1, wherein the polyA track is at least 30 nucleotides long.

6. The vector of claim 1, wherein the polyA track is at least 60 nucleotides long.

7. The vector of claim 1, wherein the IRES is from Taura syndrome virus, Triatoma virus, Theiler's encephalomyelitis virus, simian virus 40, Solenopsis invicta virus 1, Rhopalosiphum padi virus, Reticuloendotheliosis virus, human poliovirus 1, Plautia stali intestine virus, Kashmir bee virus, Human rhinovirus 2, Homalodisca coagulata virus-1, Human Immunodeficiency Virus type 1, Himetobi P virus, Hepatitis C virus, Hepatitis A virus, Hepatitis GB virus, foot and mouth disease virus, Human enterovirus 71, Equine rhinitis virus, Ectropis obliqua picorna-like virus, Encephalomyocarditis virus, Drosophila C Virus, Human coxsackievirus B3, Crucifer tobamovirus, Cricket paralysis virus, Bovine viral diarrhea virus 1, Black Queen Cell Virus, Aphid lethal paralysis virus, Avian encephalomyelitis virus, Acute bee paralysis virus, Hibiscus chlorotic ringspot virus, Classical swine fever virus, Human FGF2, Human SFTPA1, Human AML1/RUNX1, Drosophila antennapedia, Human AQP4, Human AT1R, Human BAG-1, Human BCL2, Human BiP, Human c-IAP1, Human c-myc, Human eIF4G, Mouse NDST4L, Human LEF1, Mouse HIF1 alpha, Human N-myc, Mouse Gtx, Human p27kip1, Human PDGF2/c-sis, Human p53, Human Pim-1, Mouse Rbm3, Drosophila reaper, Canine Scamper, Drosophila Ubx, Human UNR, Mouse UtrA, Human VEGF-A, Human XIAP, Drosophila hairless, S. cerevisiae TFIID, S. cerevisiae YAP1, Human c-src, Human FGF-1, Simian picornavirus, Turnip crinkle virus, or an aptamer to eIF4G.

8. The vector of claim 1, further including an RNA sequence that binds eIF4E when transcribed into the circular mRNA.

9. The vector of claim 8, wherein the RNA sequence binding to eIF4E is from Mouse histone H4, Human cyclin D1, Pea enation mosaic virus RNA2, Panicum Mosaic Virus, or an RNA aptamer to eIF4E.

10. The vector of claim 1, wherein the IRES is combined with a second IRES facilitating additional initiation factor recruitment, ribosome subunit binding, ribosome shunting, ribosome base pairing, or ribosome translocation.

11. The vector of claim 1, wherein said self-circularizing catalytic intron is a Group I intron or Group II intron.

12. The vector of claim 1, further comprising a nuclear transport element selected from Mason Pfizer Monkey Virus Constitutive Transport Element (CTE), 4E-SE element, woodchuck hepatitis virus post regulatory element, hepatitis B virus post regulatory element, or HIV rev response element.

13. The vector of claim 1, said IRES comprising SEQ ID NO. 3.

14. A method of making circular mRNA, said method comprising adding ribonucleotide triphosphates, inorganic pyrophosphatase, RNase inhibitor, and an RNA polymerase to the vector of claim 1 in appropriate reaction buffer, transcribing RNA from said vector, and allowing self-circularization of said transcribed RNA to produce circular mRNA.

15. The method of claim 14, said ribonucleotides including modified ribonucleotides m5C, m5U, m6A, s2U, Ψ, or 2′-O-methyl-U.

16. A method of making circular mRNA, said method comprising transfecting the vector of claim 1 and a phage polymerase or nucleic acid encoding a phage polymerase into a eukaryotic cell, allowing for transcription of said vector inside the cell to produce transcribed RNA, and allowing self-circularization of said transcribed RNA to produce circular mRNA.

17. A method of making circular mRNA, said method comprising transfecting the vector of claim 1 into a eukaryotic cell, wherein said vector is transcribed by a host cell RNA polymerase.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1. Shows a vector designed to produce a circular mRNA molecule. The vector is shown with an ORF inserted thereinto, but before insertion a MCS would be shown instead.

(2) FIG. 2. Shows the GFP reporter protein imaging results from circular mRNA translation in HEK cells after T7 RNA polymerase driven expression.

(3) FIG. 3. Shows the GFP and DAPI (nuclear stain) imaging results from circular mRNA translation in HEK cells after T7 RNA polymerase driven expression.

(4) FIG. 4 shows SEQ ID NOs. 1-9. SEQ ID NO. 1: T7 RNA Polymerase Promoter (21 bp); SEQ ID NO. 2: 5′ Group I Intron sequence (167 bp); SEQ ID NO. 3: eIF4E aptamer 1 sequence (86 bp); SEQ ID NO. 4: Human Beta Globin 5′ UTR (50 bp); SEQ ID NO. 5: Human Beta Globin 3′ UTR (133 bp); SEQ ID NO. 6: 3′ Group I Intron sequence (107 bp); SEQ ID NO. 7: T7 RNA Polymerase Terminator (47 bp); SEQ ID NO. 8: Polyadenylation Sequence (33 bp); SEQ ID NO. 9: EMCV IRES (593 bp).

(5) FIG. 5 shows SEQ ID NO. 10: pBSK-CR sequence=synthesized DNA sequence based on the outline provided in FIG. 5. Genes of Interest (GOI) can be cloned between NcoI and SalI and expressed as a circular mRNA with eIF4E aptamer-beta globin 5′UTR-GOI-beta globin 3′UTR-polyA.

(6) FIG. 6 shows SEQ ID NO. 11: Gene 1LG-Expresses uncapped Linear mRNA containing IRES GFP.

(7) FIG. 7 shows SEQ ID NO. 12: Gene 2CI-Expresses Circular mRNA containing IRES GFP.

(8) FIG. 8 shows SEQ ID NO. 13: Gene 3CIA-Expresses Circular mRNA containing IRES GFP, beta globin 3′UTR and polyadenylation sequence.

(9) FIG. 9 shows SEQ ID NO. 14: Gene 4EA-Expresses Circular mRNA containing eIF4E aptamer, beta globin 5′UTR, GFP, beta globin 3′UTR, and polyadenylation sequence.

(10) FIG. 10. shows one exemplary vector sequence as SEQ ID NO. 15: Circular RNA IRES-MCS-UTR-polyA-Vector contains a multiple cloning site after an EMCV IRES and before beta globin 3′UTR and polyA sequence to allow for insertion of genes in multiple reading frames. Expression yields circular mRNA.

DETAILED DESCRIPTION

(11) The current disclosure describes circular mRNA molecules that can successfully translate inside mammalian cells, as well as methods of making same, vectors for making same, and methods of using either the vector or the circular mRNA.

(12) The circular mRNA features additional regions beyond the IRES and ORF in order to help recruit ribosomes to the circular mRNA. The circular mRNA has an IRES site, an ORF for protein of interest, a 3′UTR, and an optional polyA track. In some embodiments of the invention, there can be both an IRES and a 5′ UTR, depending upon how the IRES functions. Note that given the wide diversity of IRES sequences in nature, there will be a wide range of translational efficiencies when these IRES sequences are substituted in the proposed vector. In general, however, the invention will increase the efficiency of circular mRNA product regardless of the nature of the IRES in question because of the use of the polyA tail and 3′ UTR elements, both of which help recruit ribosomes for translation.

(13) In order for circular mRNAs to translate efficiently inside cells, they must compete in vivo against cellular mRNAs also recruiting the translation machinery. While IRES sequences have fewer translation initiation requirements, the canonical recruitment process via eIF4E still recruits ribosomes much more efficiently in a head to head comparison. The present invention describes in addition to viral IRES sequences, an IRES sequence for circular mRNA that can recruit eIF4E itself.

(14) In one embodiment of the invention, the IRES is an aptamer to eIF4E (SEQ ID NO. 3) with specificity for the face of the protein that binds the guanosine cap. This aptamer has never been used as an IRES before, and represents a novel aspect of the disclosure. The aptamer is able to bind to eIF4E and thereby the rest of the translation initiation complex, similar to cellular mRNA molecules. Helping translation initiation is the presence of the 3′ UTR and polyA tail, meaning the circular mRNA in this embodiment of the invention can recruit ribosomes in almost the exact same way as cellular RNA. Cellular RNA circularizes by the way of a physical connection of PABP to eIF4G to eIF4E proteins, whereas the circular RNA in the invention is held together by a physical link.

(15) Recruitment of eIF4E for cap-independent translation can be achieved by tethering eIF4E via a peptide tag to an RNA structure that specifically binds the tag. This suggests that if an RNA structure or aptamer or stem-loop could bind eIF4E directly, than cap-independent translation could be achieved. Indeed, this has been observed already for a plant virus, Pea enation mosaic virus (PEMV). Its RNA2 molecule contains a pseudoknot RNA structure that directly binds to plant eIF4E protein.

(16) Certain eukaryotic transcripts are also able to recruit the eIF4E protein via RNA motifs in their 3′ UTRs or coding sequences, acting independently of the cap binding mechanisms. This motif is called a 4E sensitive element, or 4E-SE. Examples of 4E-SEs are found in Mouse histone H4 mRNA and Human cyclin D1 mRNA. Their role in these mRNAs is to regulate nuclear localization and export, as well as modulating translation. For the purposes of a circular mRNA, it can easily be imagined that the 4E-SE could be used to similarly recruit eIF4E to the mRNA independent of cap and stimulate translation.

(17) In another embodiment of the invention, the IRES is a 4E-SE element, taken from sequences in cellular mRNAs, which mediates direct binding to eIF4E. In this embodiment, a 5′UTR or IRES downstream of the 4E-SE may be added that promotes ribosome shunting as a way to stimulate non-canonical translation. An example would be the mouse Gtx 5′UTR or any viral or cellular IRES sequence.

(18) As an alternative to using a 4E-SE sequence to bind eIF4E, an aptamer directed against the cap-binding pocket of eIF4E is proposed to be able to replicate the effects the normal guanosine cap in promoting translation. An example eIF4E aptamer sequence is given in SEQ ID NO. 3. As is known in the art, aptamers of many degenerate sequences can be generated against a given protein, and this is only one exemplary sequence.

(19) In a similar fashion, novel IRESes could be developed that bind directly to eIF4G, skipping the necessary recruitment via eIF4E. An example of this strategy would be to develop aptamers against eIF4G that do not inhibit translation, but mediate strong binding to eIF4G.

(20) Beyond utilizing novel IRES sequences, adding other RNA elements to the circular mRNA molecule allow for translation inside cells. It is readily recognized for example, that while the cap is an important structure for eukaryotic linear mRNA translation, the 5′ UTR, 3′ UTR and polyA tails also play important roles in translation.

(21) The preferred embodiment of the invention contains a polyadenylation sequence within the circular RNA molecule of about 30-ribonucleotides of adenosine, which is able to bind to a single complex of human poly(A)-binding protein. This polyadenylation sequence would be located after the ORF, 3′UTR and before the splice site and termination signal.

(22) Polyadenylation of mRNAs have been shown to increase the expression of viral IRES driven expression. The added polyA sequence in the circular mRNA might also function as a type of additional IRES site, as suggested by a report that a polyA 5′ RNA leader could allow the bypass of initiation factors in mediating translation.

(23) In addition to viral IRESes, cellular mRNAs can also have the translation efficiency of their IRES sequences increased with polyadenylation tails. The c-myc and BiP mRNA IRES activity could be enhanced though the addition of a polyA tail, even without intact eIF4G or PABP, factors which would normally mediate such an interaction.

(24) In one embodiment of the invention, a pair of viral 5′ and 3′ UTRs may be utilized that naturally communicate with each other to mediate translation. The 5′ and 3′ UTRs of many viruses communicate through RNA-RNA or RNA-protein interactions to facilitate increased translation or regulation of translation. This suggests that optimizing the use of the said UTRs or by bringing the ends together permanently through circularization might lead to enhanced translation. One example of synergistic UTRs useful in the circular RNA invention is the pair of 5′ and 3′ UTRs from the dengue virus, which together possess IRES activity.

(25) In the invention, the ability of 3′ certain viral UTR sequences to augment or replace some of the canonical components of mRNA is also proposed. As an example, the 3′UTR of the Andes Hantavirus Small mRNA can functionally replace the polyA tail and can act in synergism with cap-dependent translation.

(26) In another embodiment of the invention, a 5′ UTR will be utilized that will facilitate the delivery of the ribosome to the first codon of the polypeptide to be translated. The mechanism of ribosomal tethering and delivery to downstream AUG codons would also be useful in circular mRNA molecules. This process is also referred to “ribosomal shunting.” An example of a sequence that mediates shunting is an mRNA element from the 5′ UTR of the Gtx homodomain mRNA, which basepairs to 18S rRNA, and the adenovirus tripartite leader.

(27) While modified RNA nucleotides have received much attention for their resistance to nucleases in the setting of siRNA among other applications, modified RNA nucleotides produce only moderate improvements in translation efficiency and transcript half-life. Thus, circular RNA represents an improvement in the ability to achieve the longest transcript half-lives compared to all other methods today, while at the same time providing a much more robust and cheaper method of mRNA production requiring only the single RNA polymerase enzyme. This is compared to other mRNA in vitro product protocols in the prior art that require up to 3 enzyme reactions total (e.g., RNA polymerase, polyadenylase, and capping enzyme). Furthermore, RNA yields from transcription reactions mixed with cap analogue are generally 2-6 times lower than without, representing another production advantage for circular mRNA.

(28) The circular mRNA described herein can also be produced in vivo inside the cell. There are two different embodiments for in vivo production of circular mRNA. In the first embodiment, DNA is delivered or integrated into nucleus. Transcription will be driven by a promoter recruiting a RNA polymerase II that is endogenous to that cell. Self-splicing would occur within the nucleus. Given that the 5′ cap has been shown to be important for mRNA export, an alternative means may need to be added in order to increase circular mRNA export. An example is the Mason Pfizer Monkey Virus constitutive transport element (CTE), an RNA sequence which helps mediate non-canonical mRNA export.

(29) The other means of in vivo circular mRNA generation would consist of transfecting linear or circular DNA containing an e.g., T7 promoter inside the cell, and adding e.g., T7. T7 polymerase protein could be transfected along with the plasmid DNA, whereafter in the cytoplasm it would bind the T7 promoter on the vector DNA and begin transcribing circular mRNA. In one embodiment of this method, the transcription cassette lacks a T7 terminator leading to continuous rolling circle transcription of RNA where T7 never dissociates from the DNA template.

(30) In other embodiments, DNA encoding T7 DNA or even T7 mRNA could be added to the cell, allowing transcription and translation to produce the T7 inside the cell. Of course, T7 is exemplary only and any similar RNA polymerase could be used, such as T6, T4, T3, SP6, or RNA Polymerase I and the like.

(31) The technologies required to produce circular RNA have been described in the literature previously. Commonly, group I self-splicing by a permuted intron-exon sequences from the T4 bacteriophage is used. This reaction can occur in prokaryotic cells, eukaryotic cells, or in vitro since it is catalyzed by RNA alone. However, a variety of different methods exist in that prior art concerning ways to synthesize circular RNA. It is understood that the proposed enhanced circular mRNA molecule could use any of these methods in its production (e.g., U.S. Pat. No. 6,210,931, U.S. Pat. No. 5,773,244).

(32) Examples of group I intron self-splicing sequences include self-splicing permuted intron-exon sequences derived from T4 bacteriophage gene td. The intervening sequence (IVS) rRNA of Tetrahymena also contains an example of a Group I intron self splicing sequences. Given the widespread existence of group I and group II catalytic introns across nature, many possible sequences could be used for creating circular RNA.

(33) Self-splicing occurs for rare introns that form a ribozyme, performing the functions of the spliceosome by RNA alone. There are three kinds of self-splicing introns, Group I, Group II and Group III. Group I and II introns perform splicing similar to the spliceosome without requiring any protein. This similarity suggests that Group I and II introns may be evolutionarily related to the spliceosome. Self-splicing may also be very ancient, and may have existed in an RNA world present before protein.

(34) Cytoplasmic expression systems have been used before as an alternative to nuclear dependent transcription, or the transfection of mRNA itself. These systems rely on the co-transfection of a phage RNA polymerase (usually T7 DNA polymerase) with a DNA template. Sometimes, the T7 is expressed as a gene from a nuclear promoter, or the mRNA encoding T7 polymerase is transfected inside the cell. These provide alternatives to protein transfection of T7 polymerase. Furthermore, T7 polymerase could direct the synthesis of more T7 polymerase in certain systems, creating a self-sustaining autogene effect. Such autogene systems achieve unparalleled expression levels, and are only limited by the amount of triphosphate-ribonucleotides in the cytoplasm among other factors.

(35) In another application of the invention, circular mRNAs could be generated continuously off a circular template, due to the highly processive nature of T7 RNA polymerase, which rarely falls off a DNA template during the elongation phase. T7 RNA polymerase can circle around plasmids many times if no proper termination sequence is provided. This has been shown in an shRNA system to produced a greatly increased yield of RNA product.

(36) In an effort to mediate translation of RNA based technologies to clinical use, advances have been made in purifying mRNA on a large scale, eliminating double-stranded RNA impurities that can activate the innate immune system (e.g., EP2510099, EP2092064).

(37) A related application distinct from circular mRNA molecules describes a circular RNA interference effector molecules (e.g., WO2010084371). Also, it has been recently published in the literature that human cells possess natural circular RNA molecules that appear to function as micro-RNA sponges. These circular RNA molecules were tested, however, and showed no translation activity, despite possessing exon sequences from proteins. In a slightly different application, the circular RNA molecules serve as substrates for Dicer and further processing to produce siRNA (e.g., EP2143792).

(38) The following experiments are exemplary only and serve to provide proof of concept experiments for the invention generally. However, the invention and the claims should not be limited by the specific exemplars provided.

Vector Construction

(39) A series of vectors were prepared to make circular mRNA matching the scheme outlined in FIG. 1. This template was then used to construct a series of different GFP encoding genes, which yield different types of mRNA molecules. The genes 1LG and 2CI produce linear mRNA molecules that exist in the prior art, while the genes 3CIA and 4CEA produce circular mRNA molecules that are novel to the current invention. As outlined below standard cloning procedures were utilized to produce the final vector DNA sequences.

(40) The plasmid, pBSK-CR was prepared with a synthesized DNA sequence matching FIG. 5 (Seq. ID No. 10). Another plasmid, pIRES-GFP, containing the EMCV IRES followed by a GFP sequence was also obtained. The following cloning steps were undertaken to produce the vectors used to generate circular mRNAs herein:

(41) Vector 1LG, which produces a linear, uncapped RNA molecule with IRES GFP-Beta Globin 3′ UTR-polyA, was constructed by digestion with BamHI and SalI in both pIRES-GFP and pBSK-CR, followed by ligation of the IRES-GFP insert into the pBSK-CR sequence.

(42) Vector 2CI, which produces a circular RNA molecule with IRES-GFP alone, was constructed by digestion of pIRES-GFP and pBSK-CR with XhoI and XbaI, followed by ligation of the IRES-GFP insert into the pBSK-CR sequence.

(43) Vector 3CIA, which produces a circular RNA molecule with IRES-GFP-Beta Globin 3′ UTR-polyA, was constructed by digestion of pIRES-GFP and pBSK-CR with XhoI and SalI, followed by ligation of the IRES-GFP insert into the pBSK-CR sequence.

(44) Vector 4CEA, which produces a circular RNA molecule with eIF4E aptamer-beta globin 5′ UTR-GFP-beta globin 3′ UTR-polyA, was constructed by digestion of pBSK-CR and pIRES-GFP with NcoI and SalI, followed by ligation of the GFP insert into the pBSK-CR sequence.

(45) Prior art plasmid pIRES-GFP produces a canonical linear capped mRNA with polyA tail that is produced inside the nucleus of a cell driven by a CMV promoter. This plasmid allows us to compare our novel circular mRNA with the expression of linear capped mRNAs and provides a direct comparison with the prior art. The relative mRNA levels produced will be different between the two systems given their different promoters, however.

Intracellular T7 Driven mRNA Expression

(46) The purpose of this experiment was to generate mRNA inside the cell with T7 polymerase, eliminating variables of toxic effects of RNA during transfection, or the possible degradation of the mRNA by abundant RNases in the environment during experimental handling. The goal was to co-transfect plasmid DNA (combinations shown below) into HEK 293 cells together with active T7 RNA polymerase protein in a 24-well format, with four wells per condition. All amounts and volumes are given on a per well basis.

(47) TABLE-US-00002 Expressed Conditions GFP expression mRNA Sequences pIRES-GFP + Lipofectamine Positive Linear capped Control IRES-GFP-polyA 1LG + T7 + Lipofectamine Positive Linear uncapped IRES- Control GFP-3′ UTR-polyA 1LG + Lipofectamine Negative None (because Control no T7 added) 2CI + T7 + Lipofectamine Test condition Circular IRES-GFP 3CIA + T7 + Lipofectamine Test condition Circular IRES- GFP-3′ UTR-polyA 4CEA + T7 + Lipofectamine Test condition Circular eIF4E aptamer- 5′ UTR-GFP-3′ UTR- polyA

(48) 1. The day before transfection, HEK cells were trypsinized and counted. Cells were plated at 1.0×10.sup.5 cells per well in 0.5 ml of complete growth medium.

(49) 2. 2 μg of DNA and 50 U of T7 RNA polymerase (NEB®) in 50 μl of serum-free OPTIMEM medium were combined, and incubated for 10 minutes at room temperature.

(50) 3. Pure lipofectamine (5 μl) was added to the plasmid/T7 RNA polymerase complex, the mixture incubated for 45 min, and then diluted to 200 μl with OPTIMEM medium.

(51) 4. After a further 30-minute incubation, 200 μl of the DNA-T7 polymerase-Lipofectamine reagent complexes were added directly to each well containing cells and mixed gently by rocking the plate back and forth. The DNA/protein/lipofectamine complexes do not have to be removed following transfection.

(52) 5. The cells are incubated at 37° C. in 5% CO.sub.2 for 24 hours.

(53) 6. Pictures of the HEK cells were then taken using fluorescent microscope at 24 hours to detect GFP expression. In some experiments, the cells were fixed at 48 hours using 4% PFA, and then stained with DAPI to detect the outline of the nucleus of cells and improve GFP visualization. Localization of GFP could then be observed in reference to the position of the nucleus.

(54) The results of the experiment showed that the linear uncapped IRES-GFP-3′ UTR-polyA expressed GFP (1LG), as has been observed in several systems. The 2CI circular mRNA, which matches the prior art of circular RNA with IRES and ORF elements only, failed to show GFP expression in repeated experiments when imaged during live cells (FIG. 2) or after cell fixation (FIG. 3). Thus, merely circularizing an RNA is not sufficient for eukaryotic expression in eukaryotic cells, even when the same mRNA is transcribable in linear form (not shown herein, but demonstrated in the prior art).

(55) The 3CIA and 4CEA circular mRNAs of the invention exhibited distinct GFP expression, which was similar in GFP intensity to the linear uncapped 1LG mRNA.

(56) TABLE-US-00003 GFP Conditions expression Sequences Results pIRES-GFP + Positive Control Linear capped +++ Lipofectamine IRES-GFP-polyA 1LG + T7 + Positive Control Linear uncapped IRES- ++ Lipofectamine GFP-3′ UTR-polyA 1LG + Negative none − Lipofectamine Control 2CI + T7 + Test condition Circular IRES-GFP − Lipofectamine 3CIA + T7 + Test condition Circular IRES-GFP- ++ Lipofectamine 3′ UTR-polyA 4CEA + T7 + Test condition Circular eIF4E aptamer- ++ Lipofectamine 5′ UTR-GFP-3′ UTR- polyA

mRNA Transcription and Transfection

(57) The same vectors 1LG, 2CI, 3CIA, and 4EA from the previous experiment were used as templates for in vitro mRNA transcription. The process of in vitro mRNA transcription is well known in the field and consists of obtaining a DNA template with a phage promoter of short length followed by the gene of interest on the same sense strand. This DNA template is oftentimes linearized due to the high processivity of RNA polymerases, but can remain circular if a polymerase terminator sequence follows after the gene.

(58) For the experiments herein, an in vitro mRNA transcription reaction was set up using the MEGAscript kit from Ambion®. A mixture of ribonucleotides, T7 polymerase and DNA template was added in a 20 μL reaction mixture. The reaction was allowed to proceed for 2 hours at 37° C. The mRNA transcripts were then purified using a standard lithium chloride protocol to remove excess ribonucleotides, DNA and protein.

(59) The purified mRNA was then transfected into HEK 293 cells using Lipofectamine, as follows:

(60) 1. The day before transfection, HEK cells were trypsinized and counted. Cells were plated at 1.0×10.sup.5 cells per well in 0.5 ml of complete growth medium.

(61) 2. 0.5-1 μg of RNA was added to 2.5 μL Lipofectamine 2000, the mixture was incubated for 45 min, and then diluted to 200 μl with OPTIMEM medium.

(62) 4. After a further 30-minute incubation, 200 μl of the mRNA-Lipofectamine Reagent complexes were added directly to each well containing cells and mixed by gently by rocking the plate back and forth. Complexes were not removed following transfection.

(63) 5. Cells were further incubated at 37° C. in 5% CO.sub.2 for 24 hours.

(64) 6. Pictures of the HEK cells were taken using fluorescent microscope at 24 hours to detect GFP expression. In some experiments, the cells were fixed at 48 hours using 4% PFA, and then stain with DAPI to detect the outline of the nucleus of cells. Localization of GFP could then be observed in reference to the position of the nucleus.

(65) The results of the mRNA transfection experiment matched the results of intracellular T7 driven mRNA expression, as expected. Linear uncapped IRES-GFP-3′ UTR-polyA mRNA expressed GFP (1LG), as has been observed in several prior art systems. The 2CI circular mRNA, which matches the prior art of EMCV IRES and ORF alone (see U.S. Pat. No. 5,766,903), failed to show GFP expression in repeated experiments when imaged during live cells or after cell fixation. Thus, consistent with the above experiments, and IRES and ORF alone are insufficient for intracellular transcription of a circular mRNA. Furthermore, it is predicted to not be sufficient for live animal (in vivo) expression either.

(66) The 3CIA and 4CEA circular mRNAs exhibited GFP expression, which was similar in GFP intensity to the linear uncapped 1LG mRNA.

(67) TABLE-US-00004 GFP Conditions expression Sequences Results pIRES-GFP + Positive Linear capped +++ Lipofectamine Control IRES-GFP-polyA 1LG + T7 + Positive Linear IRES- ++ Lipofectamine Control GFP-3′ UTR-polyA 1LG + Negative none − Lipofectamine Control 2CI + T7 + Test condition Circular IRES-GFP − Lipofectamine 3CIA + T7 + Test condition Circular IRES-GFP- ++ Lipofectamine 3′ UTR-polyA 4CEA + T7 + Test condition Circular eIF4E aptamer- ++ Lipofectamine 5′ UTR-GFP-3′ UTR-polyA

Conclusion

(68) Using two different methods of circular mRNA production, it was observed for the first time that circular mRNA can be translated intracellularly in a eukaryotic cell in direct competition with host capped mRNAs. This is a significant finding that previous researchers were unable to accomplish. Furthermore, circular mRNAs that translate inside eukaryotic cells have not been found to exist in nature so far, and thus these results are unexpected. Indeed, while circular exons and introns are now appreciated to exist inside eukaryotic cells, evolution appears not to have selected for a circular mRNA capable translation by ribosomes. A summary of the experimental results is listed in the table below.

(69) TABLE-US-00005 Vector Expressed RNA Molecule Intracellular expression pIRES- Linear capped Demonstrated in Previous Studies GFP IRES-GFP-polyA 1LG Linear IRES- Demonstrated in Current and GFP-3′ UTR-polyA Previous Studies 2CI Circular IRES-GFP No Expression in Current Study 3CIA Circular IRES-GFP- Demonstrated in Current Study 3′ UTR-polyA 4CEA Circular 4E aptamer- Demonstrated in Current Study GFP-3′UTR-polyA

(70) The gene 2CI was constructed to produce a circular mRNA molecule that matches the prior art containing the same EMCV IRES and ORF construction (see U.S. Pat. No. 5,766,903). The 2CI gene thus serves as a comparison with the current invention, which contains multiple RNA translation enhancing elements. One observes that the 2CI circular mRNA encoding EMCV IRES-GFP alone fails to produce any discernable GFP expression both in live cell imaging and after fixation inside cells. This contrasts with its reported positive expression in an in vitro rabbit reticulocyte system (Chen & Sarnow, Science, 1995).

(71) On the other hand, the circular mRNAs 3CIA and 4CEA produce expression patterns similar to the expression of linear uncapped mRNA 1LG. The linear mRNA 1LG containing uncapped EMCV IRES-GFP-beta globin 3′UTR-polyA is known in the literature to produce GFP after transfection, but we have demonstrated the first confirmed showing of expression of a circular version of the same mRNA.

(72) The difference in GFP expression between circular mRNA in 2CI (no expression) and circular mRNA in 3CIA (expression) is remarkable, considering that the only additional sequences were the beta globin 3′UTR and polyadenylation sequence. This indicates that these added sequences were able to allow the EMCV IRES to effectively recruit ribosomes inside the cell, likely through helping recruit additional initiation factors to the IRES to increase its efficiency. For example, PABP binds to polyadenylation sequence and to eIF4G, which is a targeted protein by the EMCV IRES.

(73) The present invention also describes for the first time the use of an eIF4E binding RNA sequence as an IRES-like element in recruiting ribosomes to circular RNA. So far, no mammalian viruses or cellular genes have been described that utilize eIF4E recruitment as an exclusive mechanism of ribosome recruitment. The demonstration of an eIF4E aptamer facilitating translation thus represents a novel finding for eukaryotic mRNA translation initiation.

(74) Future experiments will explore optimization of circular mRNA genes using different combinations of IRES, 5′ and 3′ UTR, and length of polyadenylation sequences. The firefly luciferase gene will be utilized as the transfected ORF to allow for quantitative measurements of protein amounts produced after mRNA translation.

(75) We also plan a future experiment to measure the half-life of our circular mRNA in eukaryotic cells, using quantitative RT-PCR and/or RNA purification and hybridization experiments. Based on the prior art teachings, we expect the half-life to be at least 2×, 3×, 4× or 5× higher than a control capped mRNA having a half-life of 10 hours Thus, we expect half-lives of at least 20 hrs, 30 hrs, 40 hrs or more.

Materials

(76) TABLE-US-00006 Reduced Serum Media Appropriate tissue culture plates and supplies T7 Polymerase (New England Biosciences ®) Lipofectamine 2000 (Invitrogen ®) HEK 293 cells maintained in Dulbecco's Modified Eagle Medium (DMEM) medium (Invitrogen ®) supplemented with 4 mM L-Glutamine (Invitrogen ®), 10% fetal bovine serum (Invitrogen ®). HEK 293 cells at 37° C. with 5% CO.sub.2. Plasmid DNA of interest Lipofectamine 2000 Reagent (store at + 4° C. until ready to use) Opti-MEM ® MEGAscript kit (Ambion ®).

(77) Each of the following references is incorporated by reference in its entirety for all purposes. EP1083232 (Nicolette); US20080299662 (Ferrandez); US20080267873 US20100047261 (Hoerr); US20100137407 (Abe); US20100173356 (Lehmann); US20110104127 (Torzewski); US20110119782 US20110143397 (Dale); US20120322864 (Rossi); U.S. Pat. No. 5,580,859 U.S. Pat. No. 5,625,047 (Felgner); U.S. Pat. No. 5,712,128 (Been); U.S. Pat. No. 5,766,903 (Sarnow); U.S. Pat. No. 5,773,244 (Ares); U.S. Pat. No. 5,824,497 (Andrews); U.S. Pat. No. 6,210,931 (Feldstein); U.S. Pat. No. 8,192,984 U.S. Pat. No. 8,257,945 (Atebekov); U.S. Pat. No. 8,278,036 (Kariko); U.S. Pat. No. 8,383,340 (Ketterer); WO2010084371 (Plane). Chen C. Y. & Sarnow P., Initiation of Protein Synthesis by the Eukaryotic Translational Apparatus on Circular RNAs. Science, New Series, 268(5209): 415-417 (1995). Culjkovic B., et al., eIF4E promotes nuclear export of cyclin D1 mRNAs via an element in the 3′ UTR., Journal of Cell Biology 169(2): 245-256 (2005). De Gregorio E., et al., Tethered-function analysis reveals that eIF4E can recruit ribosomes independent of its binding to the cap structure. RNA 7:106-113 (2001). Ford E. & Ares, M., Synthesis of circular RNA in bacteria and yeast using RNA cyclase ribozymes derived from a group I intron of phage T4. Proc. Nat. Acad. Sci. USA 91: 3117-3121 (1994). Harland R. & Misher L., Stability of RNA in developing Xenopus embryos and identification of a destabilizing sequence in TFIIIA messenger RNA. Development 102: 837-852 (1988). Konarska, M., et al., Binding of Ribosomes to Linear and Circular Forms of the 5′-Terminal Leader Fragment of Tobacco-Mosaic-Virus RNA. Eur. J. Biochem. 114: 221-227 (1981). Mackie, G. A., Stabilization of Circular rpsT mRNA Demonstrates the 5-End Dependence of RNase E Action in Vivo, Journal Of Biological Chemistry 275: 33 (2000). Mochizuki K., et al., High affinity RNA for mammalian initiation factor 4E interferes with mRNA-cap binding and inhibits translation, RNA 11: 77-89 (2005). Perriman & Ares, Circular mRNA can direct translation of extremely long repeating-sequence proteins in vivo. RNA (1998) 4:1047-1054. Perriman, R. Circular mRNA Encoding for Monomeric and Polymeric Green Fluorescent Protein. Methods in Molecular Biology 183:69-85 (2002). Puttaraju M. & Been M. D., Circular Ribozymes Generated in Escherichia coli Using Group I Self-splicing Permuted Intron-Exon Sequences, Journal Of Biological Chemistry 271(42): 26081-26087 (1996). Umekage, S. & Kikuchi Y., In vivo circular RNA production using a constitutive promoter for high-level expression. Journal of Bioscience and Bioengineering 108(4): 354-356 (2009). Umekage, S. & Kikuchi Y., In vitro and in vivo production and purification of circular RNA aptamer. Journal of Biotechnology 139: 265-272 (2009). Van Lieshout, J. F. T., et al., Ribozyme-mediated engineering of circular mRNA and its functional in vivo and in vitro translation (2007). Wang Z., et al., The cap-binding translation initiation factor, eIF4E, binds a pseudoknot in a viral cap-independent translation element. Structure 19(6): 868-80 (2011). Wang, G., et al., Cap-independent translation of human SP-A 5′-UTR variants: a double-loop structure and cis-element contribution, Am J Physiol Lung Cell Mol Physiol 296:L635-L647 (2009).